This PDF file contains the editorial “JM3 List of Reviewers” for JM3 Vol. 13 Issue 01

The great statistician and graphical expert John Tukey said, “The greatest value of a picture is when it forces us to notice what we never expected to see.”1 While many graphic forms can help us accomplish this goal, the most useful for science has proven to be the x-y scatterplot. In 2012, about 1/3 of all figures in JM3, and about 70% of all data plots, were x-y scatterplots.2 The first modern scatterplot is attributed to John Herschel (1792–1871), son of William Herschel, the discoverer of Uranus and infrared light.3 In 1833, John Herschel used a scatterplot of noisy binary star measurements to extract a trend “by bringing in the aid of the eye and hand to guide the judgment,”4 thus fulfilling Tukey’s goal. The scatterplot allows the viewer to visualize the important trends the data suggests, and possibly offer a theory to explain them, by imagining a line that passes “not through, but among them,” as Herschel so aptly said.4 By 1920, the scatterplot had come into widespread use as the tool of science we know it now to be.

This PDF file contains the editorial “Special Section on Optical Lithography Extension Beyond the 14-nm Node” for JM3 Vol. 13 Issue 01

Compact mask models provide an alternative to speed up rigorous mask diffraction computation based on electromagnetic field modeling. The high time expense of the rigorous mask models in the simulation process challenges the exploration of innovative modeling techniques to compromise accuracy and speed in the computation of the diffracted field and vectorial imaging in optical lithographic systems. The artificial neural network (ANN) approach is presented as an alternative to retrieve the spectrum of the mask layout in an accurate yet efficient way. The validity of the ANN for different illuminations, feature sizes, pitches, and shapes is investigated. The evaluation of the performance of this approach is performed by a process windows analysis, comparison of the spectra, best focus, and critical dimension through pitch. The application of various layouts demonstrates that the ANN can also be trained with different patterns to reproduce various effects such as shift of the line position, different linewidths, and line ends. Comparisons of the ANN approach with other compact models such as boundary layer model, pulses modification, spectrum correction, and pupil filtering techniques are presented.

Computational lithography, e.g., inverse lithography technique (ILT) and source mask optimization, is considered necessary for the “extremely low k1” lithography process of sub-20 nm device node. The ideal design of a curvilinear mask for computational lithography requires many changes during photomask fabrication. These range from preparation of the mask data to measurement and inspection. The manufacturability of a photomask for computational lithography is linked to predictable and manageable quality of patterning. Here, we have proposed the use of “inverse e-beam lithography” on photomask for computational lithography, which overcomes the patterning accuracy limits of conventional e-beam lithography. Furthermore, the preferred target design for ILT, a new verification method, and the accuracy required for the mask model are also discussed; with consideration of acceptable writing time (<24  h ) and computing power.

Computational lithography solutions rely upon accurate process models to faithfully represent the imaging system output for a defined set of process and design inputs. These models rely upon the accurate representation of multiple parameters associated with the scanner and the photomask. Many input variables for simulation are based upon designed or recipe-requested values or independent measurements. It is known, however, that certain measurement methodologies, while precise, can have significant inaccuracies. Additionally, there are known errors associated with the representation of certain system parameters. With shrinking total critical dimension (CD) control budgets, appropriate accounting for all sources of error becomes more important, and the cumulative consequence of input errors to the computational lithography model can become significant. In this work, we examine via simulation the impact of errors in the representation of photomask properties including CD bias, corner rounding, refractive index, thickness, and sidewall angle. The factors that are most critical to be accurately represented in the model are cataloged. CD bias values are based on state-of-the-art mask manufacturing data, and other variable changes are speculated, highlighting the need for improved metrology and communication between mask and optical proxmity correction model experts. The simulations are done by ignoring the wafer photoresist model and show the sensitivity of predictions to various model inputs associated with the mask. It is shown that the wafer simulations are very dependent upon the one-dimensional/two-dimensional representation of the mask, and for three-dimensional, the mask sidewall angle is a very sensitive factor influencing simulated wafer CD results.

Source mask optimization (SMO) is widely used to make state-of-the-art semiconductor devices in high-volume manufacturing. To realize mature SMO solutions in production, the Intelligent Illuminator, which is an illumination system on a Nikon scanner, is useful because it can provide generation of freeform sources with high fidelity to the target. Proteus SMO, which employs co-optimization method and an insertion of validation with mask three-dimensional effect and resist properties for an accurate prediction of wafer printing, can take into account the properties of Intelligent Illuminator. We investigate an impact of the source properties on the SMO to pattern of a static random access memory. Quality of a source made on the scanner compared to the SMO target is evaluated with in-situ measurement and aerial image simulation using its measurement data. Furthermore, we discuss an evaluation of a universality of the source to use it in multiple scanners with a validation and with estimated value of scanner errors.

Parallel with the introduction of EUV lithography, immersion lithography is being extended to the 14- and 10-nm node, and the lithography performance requirements need to be tightened further to enable this shrink. Next to generic scanner system improvements, application-specific solutions are needed to follow the requirements for critical dimension (CD) control and overlay. The application-specific solutions need a holistic optimization approach for the scanner, the mask, and the patterning process. We will describe the holistic lithography systems architecture that enables dynamic use of high-order scanner optimization based on advanced actuators of projection lens and scanning stages. Next to the scanner system, key components of this architecture are an angle-resolved scatterometer to measure CD, overlay, and focus, and an off-tool computation server to calculate application-specific recipes for the scanner. Based on real production wafer data, we will show the benefit for CD control, focus control, and overlay control, and demonstrate lithography performance levels required for 14- and 10-nm node production.

Although a number of factors relating to lithography and material stacking have been investigated to realize hotspot-free wafer images, hotspots are often still found on wafers. For the 22-nm technology node and beyond, the detection and repair of hotspots with lithography simulation models is extremely time-consuming. Thus, hotspots represent a critical problem that not only causes delays to process development but also represents lost business opportunities. In order to solve the time-consumption problem of hotspots, this paper proposes a new method of hotspot prevention and detection using an image recognition technique based on higher-order local autocorrelation, which is adopted to extract geometrical features from a layout pattern. To prevent hotspots, our method can generate proper verification patterns to cover the pattern variations within a chip layout to optimize the lithography conditions. Moreover, our method can realize fast hotspot detection without lithography simulation models. Obtained experimental results for hotspot prevention indicated excellent performance in terms of the similarity between generated proposed patterns and the original chip layout patterns, both geometrically and optically. Moreover, the proposed hotspot detection method could achieve turn-around time reductions of <95% for just one CPU, compared to the conventional simulation-based approach, without accuracy losses.

As 193-nm immersion lithography will likely be required to be extended beyond 40-nm half-pitch, multiple patterning lithography will become a necessity in that scenario. We present a cost-effective approach for double patterning with extendibility to sub-10-nm half-pitch division, which is a very promising candidate for advanced logic nodes. Spacers on sufficiently sloped sidewalls directly transferred from a low-contrast photoresist profile can be removed by anisotropic etching. Alternatively, spacer gaps for defining trenches may be prevented from penetrating to the substrate by the use of sloped sidewalls. These sloped sidewalls are defined by attenuated phase-shift mask features, which impart phase shifts other than 180 deg or 0 deg. Loop trimming and sidewall spacer definition are accomplished in a single photomask. In addition, there is now an extra ability to define random, arbitrary breaks in the spacer-defined pattern, without using an extra exposure for specified cuts. In this way, a single exposure using a modified attenuated phase-shift photomask, followed by a low-contrast development process around the sensitivity limit, is sufficient to pattern regularly arranged spacer-defined lines at fixed pitch while including some predetermined line cut locations.

This JM3 special section on emerging MOEMS comprises a collection of excellent papers emphasizing new technologies in MOEMS applications that have come to existence. The section includes outstanding new results in commercial research and development in photonics where micro-optics and MEMS are merged and innovative breakthrough devices come to light.

OPTRA is developing a next generation digital micromirror device (DMD) based two-band infrared scene projector (IRSP) with infinite bit depth independent of frame rate and an order of magnitude improvement in contrast over the state of the art. Traditionally, DMD-based IRSPs have offered larger format, superior uniformity, and pixel operability relative to resistive and diode arrays. However, they have been limited in contrast and also by the inherent bit depth/frame rate tradeoff imposed by pulse width modulation (PWM). OPTRA’s high dynamic range IRSP (HIDRA SP) has broken this dependency with a dynamic structured illumination solution. The HIDRA SP uses a source-conditioning DMD to impose the structured illumination on two projector DMDs—one for each spectral band. The source-conditioning DMD is operated in binary mode, and the relay optics that form the structured illumination act as a low-pass spatial filter. The structured illumination is, therefore, spatially grayscaled and more importantly is analog with no PWM. In addition, the structured illumination concentrates energy where bright objects will be projected and extinguishes energy in dark regions; the result is a significant improvement in contrast. The projector DMDs are operated with 8-bit PWM; however, the total projected image is analog with no bit depth/frame rate dependency. We describe our progress toward the development, building, and testing of a prototype HIDRA SP.

Low-cost automotive laser scanners for environmental perception are needed to enable the integration of advanced driver assistant systems into all automotive vehicle segments, which is a key to reduce the number of traffic accidents on roads. Within the scope of the European-funded project MiniFaros, partners from five different countries have been cooperating in developing a small-sized low-cost time-of-flight-based range sensor. An omnidirectional 360-deg laser scanning concept has been developed based on the combination of an omnidirectional lens and a biaxial large aperture MEMS mirror. The concept, design, fabrication, and first measurement results of a resonant biaxial 7-mm gimbal-less MEMS mirror that is electrostatically actuated by stacked vertical comb drives is described. Identical resonant frequencies of the two orthogonal axes are necessary to enable the required circle scanning capability. A tripod suspension was chosen, since it minimizes the frequency splitting of the two resonant axes. Low-mirror curvature is achieved by a thickness of the mirror of more than 500 μm. Hermetic wafer-level vacuum packaging of such large mirrors based on multiple wafer bonding has been developed to enable a large mechanical tilt angle of ±6.5  deg in each axis. Due to the large targeted tilt angle of ±15  deg and because of the MEMS mirror actuator having a diameter of 10 mm, a cavity depth of about 1.6 mm has been realized.

Digital micromirror devices (DMDs) by their high-switching speed, stability, and repeatability are promising devices for fast, reconfigurable telecommunication switches. However, their binary mirror orientation is an issue for conventional redirection of a large number of incoming ports to a similarly large number of output fibers, like with analog micro-opto electro-mechanical systems. We are presenting here the use of the DMD as a diffraction-based optical switch, where Fourier diffraction patterns are used to steer the incoming beams to any output configuration. Fourier diffraction patterns are computer-generated holograms that structure the incoming light into any shape in the output plane. This way, the light from any fiber can be redirected to any position in the output plane. The incoming light can also be split to any positions in the output plane. This technique has the potential to make an “any-to-any,” true nonblocking, optical switch with high-port count, solving some of the problems of the present technology.

To meet the high contrast requirement of 1×10⁻¹⁰ to image an Earth-like planet around a sun-like star, space telescopes equipped with coronagraphs require wavefront control systems. Deformable mirrors (DMs) are a key element of a wavefront control system, as they correct for imperfections, thermal distortions, and diffraction that would otherwise corrupt the wavefront and ruin the contrast. The goal of the CubeSat DM technology demonstration mission is to test the ability of a microelectromechanical system (MEMS) DM to perform wavefront control on-orbit on a nanosatellite platform. We consider two approaches for an MEMS DM technology demonstration payload that will fit within the mass, power, and volume constraints of a CubeSat: (1) a Michelson interferometer and (2) a Shack-Hartmann wavefront sensor. We clarify the constraints on the payload based on the resources required for supporting CubeSat subsystems drawn from subsystems that we have developed for a different CubeSat flight project. We discuss results from payload laboratory prototypes and their utility in defining mission requirements.

A method to reliably extract object profiles even with surface discontinuities that leads to 2_nπ phase jumps is proposed. The proposed method uses an amplitude-modulated Ronchi grating, which allows one to extract phase and unwrap the same with a single image. Ronchi equivalent image can be derived from modified grating image, which aids in extracting wrapped phase using Fourier transform profilometry. The amplitude of the modified grating aids in phase unwrapping. As we only need a projector that projects an amplitude-modulated grating, the proposed method allows one to extract three-dimensional profile without using full video projectors. This article also deals with noise reduction algorithms for fringe projection techniques.

The testing of a lightweight unimorph-type deformable mirror (DM) for wavefront correction in cryogenic instruments is reported. The presented mirror manufactured from the titanium alloy TiAl6V4 with a piezoelectric disk actuator was cooled to 86 K and characterized for thermally induced deformation and the achievable piezoelectric stroke between room temperature and 86 K. Through a finite element analysis, we obtained a first approximation in determining the exact temperature-dependent coefficient of thermal expansion (CTE) of the piezo material PIC151. Simulations were based on dilatometer measurements of the CTE of the TiAl6V4 mirror base between room temperature and 86 K. These investigations will enable the improvement of the athermal design of a unimorph-type DM.

Rapidly programmable micromirror arrays, such as the Texas Instruments Digital Light Processor (DLP®) digital micromirror device (DMD), have opened an exciting new arena in spectral imaging: rapidly reprogrammable, high spectral resolution, multiband spectral filters that perform spectral processing directly in the optical hardware. Such a device is created by placing a DMD at the spectral plane of an imaging spectrometer and by using it as a spectral selector that passes some wavelengths down the optical train to the final image and rejects others. Although simple in concept, realizing a truly practical DMD-based spectral filter has proved challenging. Versions described to date have been limited by the intertwining of image position and spectral propagation direction common to most imaging spectrometers, reducing these instruments to line-by-line scanning imagers rather than true spectral cameras that collect entire two-dimensional (2-D) images at once. Here, we report several optical innovations that overcome this limitation and allow us to construct full-frame programmable filters that spectrally manipulate every pixel, simultaneously and without spectral shifts, across a full 2-D image. So far, our prototype, which can be programmed either as a matched-filter imager for specific target materials or as a fully hyperspectral multiplexing Hadamard transform imager, has demonstrated over 100 programmable spectral bands while maintaining good spatial image quality. We discuss how diffraction-mediated trades between spatial and spectral resolution determine achievable performance. Finally, we describe methods for dealing with the DLP’s 2-D diffractive effects and suggest a simple modification to the DLPs that would eliminate their impact for this application.

In order to alleviate problems arising from the dynamic deformation of thin microelectromechanical systems (MEMS) micromirrors and to realize full circumferential scanning (FCS) that is highly desired in some clinic applications, such as gastrointestinal and intravascular investigations, three prototypes of polygonal pyramidal reflector-based MEMS microscanners have been developed and are described. The cascaded chevron-beam electrothermal actuator, comb-drive electrostatic actuator with double T-shaped spring mechanism, and comb-drive electrostatic actuator with resonating mechanism were investigated in detail as a means to drive a polygonal microreflector. The polygonal microreflector, which has multiple facets on its surface, was fabricated through a route involving KOH wet-etching processing and diamond-turning soft lithography technologies. The assembly process of the actuators and the microreflector is also presented. Near-FCS capability can be realized by all the three different MEMS devices. A peak scanning speed up to 180 Hz and a maximum optical scanning angle of 240 deg were achieved by the electrostatic-resonating MEMS microscanner.

Light trapping in optical cavities has many applications in optical telecommunications, biomedical optics, atomic studies, and chemical analysis. Efficient optical coupling in these cavities is an important engineering problem that affects greatly the cavity performance. Reported in-plane external fiber Fabry–Perot cavities in the literature are based on flat micromachined mirrors. In this case, the diffraction loss in the cavity is usually overcome by using an expensive-lensed fiber or by inserting a coated lens in the cavity leading to a long cavity with a small free spectral range. In this work, we report a Fabry–Perot cavity formed by a multilayer-coated cleaved-surface single-mode fiber inserted into a groove while facing a three-dimensional concave micromirror; both are fabricated by silicon micromachining. The light is trapped inside the cavity while propagating in-plane of the wafer substrate. Theoretical modeling is carried out, taking into account the effect of asymmetry in the mirror radii of curvature resulting from the micromachining process. A cavity is formed using a concave mirror with 200 and 100 μm in-plane and out-of-plane radii curvature, respectively. The presented cavity has a measured line width of 0.45 nm around 1330 nm showing a quality factor of about 3000, which resembles a one order of magnitude improvement over a flat-mirror cavity.

Hyperspectral imaging sensors have proven to be powerful tools for highly selective and sensitive chemical detection applications, but they have significant operational drawbacks including slow line-scanning acquisition, large data volume of the resulting images, and a detection time lag due to the computational overhead of the matched-filter analysis. We have recently developed and demonstrated a high-speed, high-resolution, programmable spectral filter based on the Texas Instruments DLP® digital micromirror device (DMD) that is capable of performing matched-filter image processing across a two-dimensional (2-D) field-of-view directly in the optical hardware and will enable real-time chemical detection without slow scanning, large data volumes or expensive postprocessing requirements. Based on traditional optical techniques, our spectral filter encodes the spectral information orthogonal to the spatial image information, enabling the DMD to encode matched-filter information into the spectral content of the scene without disturbing the underlying 2-D image. With this new technology, everything from simple multiband filters to very complicated hyperspectral matched filters can be implemented directly in the optical train of the sensor, producing an image highlighting a target signature within a spectrally cluttered scene in real time without further processing. We will first describe the implementation of a DMD as a multiplexing spectral selector for a 2-D field-of-view, discuss its utility as a multiband spectral filter, and show how the DLP’s® duty cycle-based grayscale capability enables the direct measurement of the adaptive matched filter. We will also show examples of multispectral and hyperspectral matched-filter images recorded with our visible spectrum prototype.

A new design concept for a dynamically focusing silicon membrane mirror with 6-mm diameter and electrostatic actuation was realized. With this concept, membrane buckling by residual compressive stress inside the membrane can be avoided, which is observed even in crystalline membranes fabricated in silicon-on-insulator (SOI) technology and leads to severe distortion of stress-sensitive devices, such as membrane-based micromirrors. To eliminate the influence of residual stress (compressive or tensile), a membrane suspension with a new stress-relief design was developed by the use of finite element (FEM) simulations. The improvement was achieved by a special tangential-beam suspension, which allows an in-plane expansion or contraction of the membrane, which reduces the stress-induced deformation and leads to substantially flat and distortion-free micromirrors (distortion<λ/10 ). Measurements of realized devices are in very good agreement with the prediction of the FEM simulation. A comparison of membranes with the new stress-relief suspension shows, for example, for a membrane with 6-mm diameter and 10-μm thickness, a distortion of 54 nm compared to 340 nm for a conventional rigidly clamped membrane. A focus range between 97 mm and infinity (flat position) can be used in accordance with the simulation.

Spatial light modulators (SLMs) have shown to be versatile tools for displays, but also for various applications in optical metrology since light can be individually directed and customized and thus they may serve as flexible masks or holograms. In contrast to SLMs based on liquid crystal on silicon or liquid crystal displays (LCDs), which allow tailoring phase and/or amplitude of a wavefront in transmission or reflection, digital micromirror devices (DMDs) offer highest reflectivity and allow precise as well as fastest guidance of light rays based on the fundamental reflection law. Here, we present and compare different approaches for simulating the diffractive pattern when applying a micromirror device for wavefront readout. The different simulation methods for calculating the diffraction pattern are based on Monte-Carlo simulations in combination with nonsequential ray tracing, on Fourier optics methods (Fourier transform, FT) and on complex digital holographic wavefront propagation. The wavefront measurement concept with the DMDs is based on selecting single subapertures of the wavefront under test and on measuring the wavefront slope consecutively in a scanning procedure. In contrast, in LCD-based approaches already shown in literature, the selection of subapertures and thus the scanning procedure is performed in transmission. The measurement concept and diffraction-based measurement errors of this method will be demonstrated for aspheric optics. Furthermore, different approaches for the prediction and reduction of the diffractive pattern—also based on holographic complex wave front propagation—will be described and characterized.

This paper reports on a gimbaled MEMS scanning mirror with quasistatic resonant actuation, specially developed for adaptive raster scanning in an innovative three-dimensional (3-D) time-of-flight (ToF) laser camera with real-time foveation. Large quasistatic deflections of ±10  deg are provided by vertical comb drives in the vertical direction in contrast to resonant horizontal scanning. This mirror is 2.6×3.6  mm and operates at 1600 Hz with an 80-deg optical scan range. For position feedback, piezo-resistive position sensors are integrated on chip for both axes. To guarantee the full reception aperture of effectively 5 mm, a synchronized driven MEMS scanner array—consisting of five hybrid assembled MEMS devices—is used in an innovative 3-D ToF laser scanner. This enables a distance measuring rate of 1 MVoxel/s with an uncertainty in distance measurement of 3 to 5 mm for a 7.5-m measuring range for a gray target. Flatness-based open loop control is used for driving control of quasistatic axis in order to compensate for the dynamics of the low damped MEMS system.

A translatory MOEMS actuator with extraordinary large stroke—especially developed for fast optical path length modulation in miniaturized Fourier transform infrared spectrometers (FTSs)—is presented. A precise translational out-of-plane oscillation at 500 Hz with large stroke of up to 1.2 mm is realized by means of an optimized MEMS design using four pantograph suspensions of the comparative large mirror plate with 5-mm diameter. The MOEMS device is driven electrostatically resonant and is manufactured in a CMOS compatible silicon-on-insulator process. Up to ±600  μm amplitude (typically 1 mm stroke) has been measured in vacuum of 30 Pa and 50 V driving voltage for an optimized pantograph design enabling reduced gas damping and higher driving efficiency. For FTS system integration, the MOEMS actuator has been encapsulated in a hybrid optical vacuum package. The thermal influences of packaging technology on MOEMS behavior are discussed in more detail.

Hyperspectral infrared imagers are of great interest in applications requiring remote identification of complex chemical agents. The combination of mercury cadmium telluride detectors and Fabry–Perot filters (FPFs) is highly desirable for hyperspectral detection over a broad wavelength range. The geometries of distributed Bragg reflector (DBR)-based tunable FPFs are modeled to achieve a desired spectral resolution and wavelength range. Additionally, acceptable fabrication tolerances are determined by modeling the spectral performance of the FPFs as a function of DBR surface roughness and membrane curvature. These fabrication nonidealities are then mitigated by developing an optimized DBR process flow yielding high-performance FPF cavities suitable for integration with hyperspectral imagers.

Three-dimensional integration of stacked device cells in front end (FE) and of advanced metallization in back end (BE) enabled by through-silicon via (TSV) opens new paths to increased product functionality even without device shrink. But “going 3-D” also creates many new challenges for the semiconductor industry. Development of the new designs, manufacturing methods, and processes demands rapid technology and materials characterization, as well as in-line metrology and inspection for process development and control in completely new and changing applications environments.

Today three-dimensional system-in-package integration together with advanced interconnect technologies based on through silicon vias, through encapsulant vias, and microbumps are considered some of the most promising enabling technologies for “More than Moore” solutions. These technologies involve vertical die stacking or chip embedding with high-density interconnects and are based on combinations of process steps that come from formerly strictly separated technology areas. Thus, there is an increasing need to understand a large number of different interface properties, to control and optimize processes, and to avoid defect formation that could affect reliability. This complexity, in terms of design, new materials, and material combinations, also requires the development of new failure analysis tools to support these developments. The application potential of a new fast plasma focused ion beam (FIB) system for metrology and failure analysis is demonstrated in several selected case studies. The higher material removal rate of this system improves the range of application fields and/or the analysis throughput. This makes the plasma-FIB a very attractive tool for the analysis of relatively large interconnect structures without any need for mechanical preparation steps.

Potential challenges with managing mechanical stress and the consequent effects on device performance for advanced three-dimensional (3-D) IC technologies are outlined. The growing need in a simulation-based design verification flow capable of analyzing a design of 3-D IC stacks and detecting across-die out-of-spec variations in MOSFET electrical characteristics caused by the die thinning and stacking-induced mechanical stress is addressed. The development of a multiscale simulation methodology for managing mechanical stresses during a sequence of designs of 3-D IC dies, stacks, and packages is focused. A set of physics-based compact models for a multiscale simulation is proposed to assess the mechanical stress across the device layers in silicon chips stacked and packaged with the 2.5D interposer-based, and true 3-D through silicon via-based technology. A simulation flow is developed for the hot-spot checking in different types of devices/circuits such as digital, analog, analog matching, memory, IO, characterized by different sensitivities to the stress-induced mobility variations. A calibration technique based on fitting to measured electrical characteristics of the test-chip devices is presented. The limited characterization or measurement capabilities for 3-D IC stacks and a strict “good die” requirement make this type of analysis critical in order to achieve an acceptable level of functional and parametric yield.

We are developing a new macroinspection technology for through silicon via (TSV) process wafers. We present new simulation results obtained with a fine TSV model and new optics. The optical system includes not only diffraction optics, but also polarization optics, by which we can detect changes in the profile (cross-sectional shape) of repeated patterns by detecting changes in the polarization status of reflected light. We confirmed the performance of the methodology by optical simulation using a model of via patterns with 1 μm diameter and 10 μm depth as a typical intermediate-interconnect-level TSV.

Characterization of silicon stress near copper (Cu)-filled through-silicon via(s) (TSVs) was demonstrated using high-resolution micro-Raman spectroscopy. For depth profiling of Si stress distribution near TSVs, a polychromator-based, multiwavelength excitation Raman measurement with different probing depths was used. The design concept of the polychromator-based, multiwavelength micro-Raman spectroscopy system, including the importance of the high-spectral resolution and multiwavelength excitation capability in three-dimensional (3-D) Si stress characterization, was described. Silicon stress near Cu-filled TSVs, with various sizes and layouts, was measured and analyzed before and after Cu annealing steps. Main factors impacting Si stress near Cu-filled TSVs are analyzed based on Raman characterization results on various types of TSV structures, layouts, and Cu annealing conditions. Large variations in Si stress in TSV arrays were measured in wafers with poor Cu fill characteristics and in wafers annealed in nonoptimized conditions. The Cu annealing sequence and annealing conditions are found to be significantly important for managing Si stress and the reliability of Cu-filled TSVs. Substantially lower Si stress was measured near Cu-filled TSVs with voids. Multiwavelength micro-Raman spectroscopy can be used as a very effective noncontact, nondestructive, inline TSV process and Si stress monitoring technique.

This paper focuses on the metrology needs and challenges of through-silicon via (TSV) fabrication, consisting of TSV etch, liner, barrier, and seed (L/B/S) depositions, copper plating, and copper chemical mechanical planarization. These TSVs, with typical dimensions within a factor of two or so of ≈5  μm  ×50  μm (diameter×depth ), present an innovative set of metrology challenges because of the high aspect ratio and large feature sizes. The metallization deposition process includes thin layers of L/B/S metal; metrology for these layers determines whether there is complete coverage of the sidewalls. Metrology for the fill step includes verifying that the TSVs are deposited without voids and that the extent of stress on the surrounding silicon does not exceed acceptable limits.

Current trends in microelectronics focus on three-dimensionally integrating different components to allow for increasing density and functionality of integrated systems. Concepts pursued involve vertical stacking and interconnecting technologies that employ micro bumping, wafer bonding, and through silicon vias (TSVs). Both the increasing complexity and the miniaturization of key elements in three-dimensional (3-D) components lead to new requirements on inspection and metrology tools and techniques as well as for failure analysis methodologies. For metrology and quality assessment in particular, methods operating nondestructively are of major importance. Scanning acoustic microscopy has the ability of illuminating optically opaque materials and, thus, allowing the assessment and imaging of internal structures. Conventional scanning acoustic microscopy (SAM) equipment can be applied to analyze the quality of wafer-bonded interfaces in 3-D integration but may reach its limitations when structures shrink in size and gain complexity. A new concept of acoustic inspection in the gigahertz (GHz) frequency band is explored for its applicability to 3-D integration technologies. Extending the acoustic inspection frequency allows for lateral resolutions in the 1-μm range and also enables the inspection of microbumps and TSVs in addition to wafer bonded interfaces, which exceed the applicability of conventional SAM. Three case studies are presented here ranging from conventional SAM on a full wafer scale to acoustic GHz microscopy on thin films and TSVs.

The three-dimensional (3-D) integrated circuit relies on the stacking of multiple two-dimensional integrated circuits into a single device using through silicon vias (TSVs) as the vertical interconnect. There are a number of factors driving 3-D integration, including reduced power consumption, resistance–capacitance delay, form factor, as well as increased bandwidth. One of the critical process steps in all 3-D processes is stacking, which may take the form of wafer-to-wafer, chip-to-wafer, or chip-to-chip bonding. This bonding may be temporary, such as can be used for attaching a device wafer to a handle wafer for thinning, or permanent, incorporating direct metal bonds or solder bumps to carry signals between the wafers and oxide bonds or underfill in the regions without conductors. In each of these processes, it is critical that the bonding is executed in such a way to prevent the occurrence of voids between the layers. This article describes the capabilities of infrared (IR) microscopy to detect micrometer size voids that can form in optically transparent blanket media such as oxide-to-oxide permanent bonding, benzocyclobuten permanent bonding, or temporary adhesive bonding laminate interfaces. The infrared microscope is described, and the measurement results from a bonded void wafer set are included. The wafers used to demonstrate the tool’s capabilities include programmed voids with various sizes, densities, and depths. The results obtained from the IR microscopy measurements give an overview of the technique’s capability to detect and measure voids as well as some of its limitations.

The technology for semiconductor device packaging is rapidly advancing in response to increasing demand for smaller and thinner electronic devices. Three-dimensional chip stacking that uses through-silicon vias (TSVs) is a key area of technical focus, and the high-density TSV (HDTSV) is a major enabler of three-dimensional (3-D) integrated circuit technology. The ongoing development of this novel technology has created a need for noncontact characterization. One of the main challenges for 3-D TSV metrology is measuring high aspect ratio features that limit conventional optical microscopy techniques. We demonstrate the use and enhancement of an existing wafer metrology tool, a spectral reflectometer, by developing and implementing theoretical models and measurement algorithms for inspection of HDTSVs. It is capable of measuring the depth of vias, and can also be used for the estimation of bottom roughness and bottom shapes of vias through the model fitting. Our nondestructive solution has measured TSV diameters as small as 3 μm and aspect ratios <16.5∶1 . Submicron depth measurement accuracy has been verified in the range of 30 to 60 μm on via depth. Metrology results from actual wafers formed from 3-D interconnect processing are presented.

The angles of diffraction of a grating are a function of the incident wavelength and the period of the grating. The ability to vary the period of a grating is of interest in a variety of applications, as this in effect provides tunability of the diffraction angles. Most tunable gratings vary the period in an analog fashion. This is achieved by pulling the grating beams apart, which varies the gap between the beams, and hence the period. Changing the period in this way usually does not alter the width of each beam. Although this provides tunability with a high resolution, the ratio of the width of each grating beam to the period changes for each period. This in turn changes the efficiency for each grating period. In this work, we present a design of a MOEMS grating that changes the period between two values, i.e., from 12 to 24 μm and 48 μm, while keeping the efficiency constant. Electrostatic attraction is used to selectively pull down beams in order to achieve this. Simulation and experimental results show a variation of about 1% in efficiency between the two grating states. The measured efficiencies in the first order of the 24- and 48-μm period gratings were found to be 6.5% and 7% compared to the simulated values of 9% and 10%, respectively. The gratings were fabricated using surface micromachining. The process flow is presented.

A microfabricated hollow air-filled waveguide with two back-to-back bends operating in the WR-3 frequency range of 220 to 325 GHz has been successfully fabricated through a multilayered ultrathick SU8 process and has ultralow transmission loss of 0.028 to 0.03  dB/mm in a performance test. This is the first SU8 constructed hollow waveguide operating at WR-3 frequencies having transmission loss comparable with the state-of-the-art computer numerical control (CNC) machined metal circuits. The multilayered SU8 processing technique has a great advantage of reducing the transmission loss by decreasing the probability of air gaps between SU8 layers. The excellent device performance is also attributed to the precisely controlled layer thickness and ultimately low surface roughness of evaporated silver which only contributes a small part of the microwave signal attenuation.

The growth of vertically self-aligned ZnO nanorods in an anodic aluminum oxide (AAO) template is investigated, along with their wetting characteristics. The highly ordered through-hole AAO template was prepared from an aluminum foil of 99.95% purity in a mixed solution of phosphoric and oxalic acids at 160 V at 1°C, and the resulting pore diameter and channel length are 250 nm and 2 μm, respectively. The vertically self-aligned ZnO nanorods produced with AAO template show the wetting transition time from hydrophobic to superhydrophilic under UV illumination in 9 min. This fast transition time is attributed to the high surface area to volume ratio of the ZnO nanorods and the adsorption enhancement of water molecules by UV-generated electron-hole pairs.

An optimized detector fabrication process is developed for resonator-quantum well infrared photodetectors (R-QWIPs). The R-QWIPs are the next generation of QWIP detectors that use resonances to increase the quantum efficiency (QE). To achieve the expected performance, the detector geometry must be produced in precise specification. In particular, the height of the diffractive elements and the thickness of the active resonator must be uniformly and accurately realized to within 0.05-μm accuracy. To achieve this specification, an optimized inductively coupled plasma etching process with a nearly infinite etching selectivity for the GaAs over the Al_x Ga_1−xAs etch-stop layer was developed. Using this etching technique, we have fabricated a number of R-QWIP test detectors with the required dimensions. Their QE spectra were tested to be in close agreement with the QE predictions.

A comparison of dry and wet release methods for surface micromachining of metallic structures, such as RF MEMS switches, test structures, bridges, and cantilevers is presented. The dry release process is optimized by varying the concentration of O ₂ and CF ₄ plasma and RF power. The plasma ashing of the sacrificial layer typically results in damage to metallic structures or stress-related deformation due to rise in temperature (<80°C ). A wet release process using critical point drying (CPD) has been investigated to realize gold-electroplated structures with reduced residual stress. The CPD, being a low-temperature (31.1°C) process, is more suitable for compliant structures without any deformation.

The temperature rise from the operation of implanted biomedical devices should be kept within a safe limit to prevent thermal damage or any undesirable thermal effects. To evaluate the temperature rise from the operation of an implanted high-density microelectrodes array (MEA) on a flex in the subretinal space, we directly integrated resistance temperature detectors into this retinal MEA device in the same micro fabrication. We surgically implanted this MEA device in the subretinal space of a rabbit model and measured the temperature rise in vivo. The measured temperature rise is consistent with the calculated values from the finite element method. Experiment showed the temperature versus power dissipation of the MEA device had a slope of 0.84°C/(mW⋅mm ⁻² ) . To ensure the temperature rise is within 1.0°C, the maximum power dissipation on the retinal implant should be kept within 1.2  mW⋅mm ⁻² .

For the direct measurement of microgripper gripping force and the calibration of microgripper jaw stiffness, this paper presents a new microforce measuring device based on low-cost SU-8 microcantilever sensors with integrated copper piezoresistive strain gauge. On the basis of the deduced equations, the geometric parameters of the microcantilever sensor are determined. Then, the fabrication of the sensors is carried out using a simple process. One fabricated sensor is calibrated by measuring the output voltage and the applied vertical force simultaneously. The calibration result of the microcantilever stiffness is 2.99  N/m and the output sensitivity is 0.51  V/N . The performance test results of the calibrated sensor show that the force sensing range is 405 μN and the maximum nonlinearity is 11 μN. Finally, the microgripping forces of two different microgrippers are measured by the developed device, and jaw stiffness calibrations are also carried out. According to the experiment results, the normally closed SU-8 microgripper has a jaw stiffness of ∼2.83  N/m and the jaw stiffness of the normally open SU-8 microgripper is ∼7.22  N/m .

Excimer laser at 248 nm has been used to micromachine holes and channels on stainless steel and polymethyl methacrylate using mask projection methods. The machining is numerically simulated considering the laser beam as a heat source only, thermal diffusion, and melt vaporization. The depth and width of the machined features at different pulse energies and number of pulses are measured by optical profilometry. Both the depth and the aspect ratio (depth-to-width) are found to increase with increasing number of pulses and pulse energies. The depth predicted by the simulations based on the thermal ablation model is closer to the experimental observations for steel as compared with that for the polymer. This allows for the quantification of the contribution of thermal ablation processes in the excimer laser machining of metals and polymers. High-energy pulses are found to have higher ablation efficiency in case of metal and lower ablation efficiency in case of polymer.

High-sensitivity extreme ultraviolet (EUV) mask pattern defect detection is one of the major issues remaining to be addressed in device fabrication using extreme ultraviolet lithography. In order to achieve inspection sensitivity and suitability for the 1× nm node, a projection electron microscope (PEM) system is employed that enables high-speed/high-resolution inspection, which is not possible using conventional deep ultraviolet or electron beam inspection systems. By employing higher electron energy in the electron optics (EO) exposure system and by improving the PEM design, we have minimized the aberration that occurs when working with EO systems and we have improved the transmittance of the system. Experimental results showing the improved transmittance were obtained by making electron throughput measurements. To guarantee the tool’s aptness for 16-nm node EUV mask inspection, corresponding sized programmed defects on masks were designed, and the defect detection sensitivity of the EO system was evaluated. Improvements in image resolution and electron throughput have enabled us to detect 16-nm sized defects. The PEM system was integrated into a pattern inspection system for defect detection sensitivity evaluation.

We present a method to etch vertical and smooth silicon sidewalls. The so-called modified Bosch process combines a passivation step, a breakthrough step, and an anisotropic etch step within one repeatable loop. Thanks to the sidewall protection from the passivation step and the anisotropic nature of the etch step, this method can etch smoother silicon sidewalls than either a typical Bosch process or a continuous anisotropic etch process. The silicon sidewall of a 4-μm-deep waveguide structure etched by this method has a root mean square roughness below 10 nm and a peak-to-valley (P-V) roughness below 60 nm. Comparisons between silicon waveguides etched by this process and by another deep reactive-ion etching process showed remarkable improvement in propagation loss at the wavelength of 1550 nm.

A major challenge for extreme ultraviolet (EUV) lithography is avoiding defects in the fabrication of multilayered (ML) mask blanks. Substrate defects and adders during ML coating are responsible for ML defects, which cause changes on phase and amplitude of EUV light. The ML defects must be identified by inspection prior to absorber patterning in order to reduce the effects of ML defects via covering them with patterns to permit the use of fewer ML defect blanks. Fiducial marks (FMs) on ML blanks can be used for mask alignment and to accurately and precisely determine the locations of ML defects. In this study, we fabricated an FM mask by resist exposure using an e-beam writer and etching. Then, we inspected FMs and ML defects with an EUV actinic full-field mask blank inspection tool developed by EIDEC-LaserTec (LT ABI; EIDEC, Tsukuba, Japan and LaserTec, Yokohama, Japan). Next, we evaluated the ML defect location accuracy on the mask based on FMs of several line depths. Here, we explain the estimation of defect location accuracy using FMs and the LT ABI and discuss the defect numbers which can be covered by absorber patterns. Fewer than 19 defects per blank should be required for EUV blanks to cover the ML defects with patterns.

A lattice-type Monte Carlo–based mesoscale model and simulation of the lithography process have been adapted to study the insoluble particle generation that arises from statistically improbable events. These events occur when there is a connected pathway of soluble material that envelops a volume of insoluble material due to fluctuations in the deprotection profile. The simulation shows that development erodes the insoluble material into the developer stream and produces a cavity on the line edge that can be far larger than a single polymer molecule. The insoluble particles can coalesce to form aggregates that deposit on the wafer surface. The effect of the resist formulation, exposure, postexposure bake, and development variables on particle generation was analyzed in both low- and high-frequency domains. It is suggested that different mechanisms are dominant for the formation of line-edge roughness (LER) at different frequencies. The simulations were used to assess the commonly proposed measures to reduce LER such as the use of low molecular weight polymers, addition of quenchers, varying acid diffusion length, etc. The simulation can be used to help set process variables to minimize the extent of particle generation and LER.

With the rapid development and technical support by microelectromechanical systems, a lab-on-a-chip or micrototal analysis system has been widely developed for medical tests due to its low cost, simple, disposable, and less contamination. Comparing to the continuous microfluidics system, a digital microfluidics system has the advantages of being without micropump and easily driven using the alternative change of hydrophilic and hydrophobic properties. A feedback control system is proposed for driving digital microfluidics by using capacitive catenary as a sensor. In the proposed system, the moving droplet can be real-time positioned and controlled to overcome the imperfection of fabricated devices and the environmental disturbance. Furthermore, with the compensation of adjusting the amount of applied voltage online, the droplets can run smoothly toward the desired position in the chip through the feedback control. As a result, the proposed system is capable of moving droplets with higher speed, compared to most of the existing system, and improves the reliability of handling microfluidics. The chip for microfluidics operation has been successfully fabricated. With the self-designed circuit and LabVIEW interface, the experiments for operating microdroplets in an open channel are conducted to verify the performances of proposed devices.

The thermal stability of La/B and LaN/B multilayers was investigated. The two multilayer systems were found to have comparable subångström period expansion upon annealing in the temperature range of 250°C to 400°C. For La/B multilayers, wide angle x-ray diffraction analysis revealed that the size of LaB ₆ crystallites present did not change significantly upon thermal treatment. Using grazing incidence x-ray reflectometry, strong change in the internal structure due to interdiffusion at the interfaces of La/B multilayers was observed after annealing. This, coupled to the unchanged crystallinity, suggested the growth of amorphous lanthanum boride interlayers. At wavelength reflectance, measurements showed that as-deposited LaN/B multilayers had an enhanced optical contrast compared with La/B. During thermal loading, the rate of diffusion-induced reflectance decrease in LaN/B multilayers was slower than in La/B. The enhanced thermal stability of LaN/B was attributed to the slower growth of LaN-B interfaces compared with La-B.

Optimization technologies have been widely applied to improve lithography performance, such as optical proximity correction and source mask optimization (SMO). However, most published optimization technologies were performed under fixed process conditions, and only a few parameters were optimized. A method for mask, process, and lithography-tool parameter co-optimization (MPLCO) is developed to extend the process window. A normalized conjugate gradient algorithm is proposed to improve the convergence efficiency of the MPLCO when optimizing different scale parameters. In addition, a parametric mask and source are used in the MPLCO that could obtain exceedingly low mask and source complexity compared with a traditional SMO.

This paper presents a structural improvement method for temperature coefficient of resonance frequency (TCF) in resonant silicon sensors. A silicon resonator, whose mass was suspended by a slanted flexible beam, was adopted in this study. The slanted suspension beam was formed by (1 0 0) and (1 1 1) crystal planes and fabricated by anisotropic wet etching. We propose a stress buffer structure to improve the robustness of resonance frequency against temperature variations. Theoretical considerations of the tested resonator are proposed to augment the effect of the buffer structure. The temperature dependence of the resonance frequency is experimentally characterized over the range −40°C to 60°C. The TCF of the original resonators with no stress buffer structure was linearly fitted to be 36 and 40  ppm/°C . After using an appropriate stress buffer structure, the TCF is linearly fitted to be −0.98 and 0.36  ppm/°C . The experimental results suggest that the TCF of the resonator is improved to sub-ppm/°C level by using a stress buffer structure, which has more than an order of magnitude improvement comparing to the original one. The small range of TCF is much more convenient to be compensated by electrical ways.

We propose an analytical framework to build a microfluidic microsphere-trap array device that enables simultaneous, efficient, and accurate screening of multiple biological targets in a single microfluidic channel. By optimizing the traps’ geometric parameters, the trap arrays in the channel of the device can immobilize microspheres of different sizes at different regions, obeying hydrodynamically engineered trapping mechanism. Different biomolecules can be captured by the ligands on the surfaces of microspheres of different sizes. They are thus detected according to the microspheres’ positions (position encoding), which simplifies screening and avoids target identification errors. To demonstrate the proposition, we build a device for simultaneous detection of two target types by trapping microspheres of two sizes. We evaluate the device performance using finite element fluidic dynamics simulations and microsphere-trapping experiments. These results validate that the device efficiently achieves position encoding of the two-sized microspheres with few fluidic errors, providing the promise to utilize our framework to build devices for simultaneous detection of more targets. We also envision utilizing the device to separate, sort, or enumerate cells, such as circulating tumor cells and blood cells, based on cell size and deformability. Therefore, the device is promising to become a cost-effective and point-of-care miniaturized disease diagnostic tool.

This article [J. Micro/Nanolith. MEMS MOEMS. , 12, (4 ), 041203 (2013)] was originally published online on 25 September 2013 with errors in Figure 6. The figure sequence in the caption did not match the sequence given in the figure. Also, Fig. 6(d) showed an image of aluminum mask with a crossbar on the image which is not visible and an improper scale with a typographical error. The corrected figure and caption appear below. Online versions were corrected on 08 January 2014.